Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Abstract—We propose a new routing protocol BPMR
(Battery Power Management Routing) for MANET (Mobile Ad
Hoc NETwork). This protocol maintains high availability of
network nodes. The duration time of participating to a MANET
differs among nodes. Each node must work as a relay node
during the participation and consume its battery. It is not fair
that the nodes participating to the MANET for a long time relay
many packets and the ones participating for a short period
relay few packets. MANET protocols should reduce the relay
burden to the nodes that have contributed to the MANET for a
long time. MBCR (Minimum Battery Cost Routing) was
proposed for MANET, in which the route selection considers
each node’s remaining battery in order to improve the
availability of MANET. In order to improve the availability, the
nodes must not fall out early because of the battery exhaust. We
propose BPMR which considers both of the participation
duration and the remaining battery for the path selection.
BPMR achieves fairness among nodes and high availability.
The total remaining battery capacity by BPMR is superior to
the one by MBCR. We show the effectiveness of BPMR by
network simulations.
Index Terms—Battery consumption, fairness, MANET,
MBCR, participation duration, routing protocol.
I. INTRODUCTION
Because of recent advances of the wireless network
technology, there are wireless devices everywhere in our
daily life. Most networks are based on an infrastructure that
uses access points as wireless communication base stations.
MANET (Mobile Ad Hoc NETwork) [1] does not require
such wireless communication base stations. It is expected as a
new type of wireless network. The paper [1] insists that
MANET will see widespread use within the next few years.
MANET is an independent network in which devices
communicate only by their equipped wireless
communication modules. MANET is expected to be used as a
temporary network for mobile phones when the cellular
network infrastructure is not available in the case of a disaster.
It is also expected to be used in the mountain areas where
there are no fixed-line networks. Devices within the scope of
their radio waves can directly communicate with each other.
In addition, each device has a function as a router. Thus a
device can communicate to a device outside the scope of its
radio wave, because the devices can relay the other devices’
packets. Wireless devices often move in MANET. Thus the
network topology changes with time. In MANET, the route
to a destination device changes frequently because of the
Manuscript received September 30, 2015; revised January 20, 2016. This
work was supported by JSPS KAKENHI Grant Number 26330019.
The authors are with Kogakuin University, Tokyo, 163-8677 Japan (e-mail: [email protected], [email protected]).
movement of the devices. Thus the routing protocols for
MANET need to reconstruct the routing paths soon when the
network topology changes.
Another usage of MANET is avoiding wireless network
congestion. The traffic of cellular networks rapidly increases
by mobile phones, thus the communication speed decreases.
To avoid the decrease, load balancing is necessary. Currently,
Wi-Fi spots are operated near many train stations for the
purpose. However, setting up Wi-Fi spots is not easy and
Wi-Fi has a disadvantage of its narrow effective radio range.
Thus, there are areas where the load balancing cannot be
easily performed. In order to solve the problem, short range
communication among the users within the area can be
performed by MANET, thus only long distance
communication uses the infrastructure and load valancing is
achieved. These types of works [2], [3] aim to improve the
throughput. The paper [2] tried to reduce collisions for
throughput improvement (MAC layer) in MANET. The
authors showed the improvement by a simulation. The paper
[3] compared famous protocols (DSDV, DSR and AODV)
from the point of the number of packets, throughput, and
delay by a simulation. Authors insisted that the best protocol
is DSR.
Some protocols for MANET have been proposed to
achieve low power consumption [4]-[6], however very few
works have been done for the following fair load valancing
problem.
In current MANET protocols, every device has the same
probability to be selected as a relay device at the path
selection. It is not fair that the devices participating to the
MANET for a long time relays many packets and the ones
participating for a short period relay few packets. Since
devices wish for high speed communication and the
minimum delay, high performance devices tend to be
selected as the relay devices. Thus, these devices use more
battery power than the other devices. Since the battery
capacity of devices is limited, heavy load devices fall out
from the network within a short period. It results in
decreasing the success rate of communication and throughput
of the network. Thus, we need a new protocol that aims long
term communication stability. To achieve the purpose, the
route selection must consider power consumption.
There are several protocols [7], [8] for MANET that
considers the power consumption. MBCR [7] avoids to use
low remaining battery devices, thus such devices tend not to
fall out from the network early. However, MBCR sometimes
increases the total power consumption in order to keep
devices alive. It is important to keep devices alive in MANET
but increasing the total power consumption is not desirable.
DSR [8] was designed as a reactive type routing protocol for
MANET. This protocol is one of the most protocols for
Battery Power Management Routing Considering
Participation Duration for Mobile Ad Hoc Networks
Masaru Yoshimachi and Yoshifumi Manabe
Journal of Advances in Computer Networks, Vol. 4, No. 1, March 2016
13doi: 10.18178/jacn.2016.4.1.196
MANET and registered as RFC 4728. DSR is simple and
efficient. However it has some problems in the security,
power consumption, fairness, and so on because of its
simpleness. The paper [4] indicated that the one of the most
important issue of routing protocols is to provide energy
efficient routes and the transmission power control is
necessary. The optimal adjustment of the power level is
essential not only for energy conservation but also for the
interference control. The paper [6] insisted that adjusting the
transmit powers of nodes in a multi hop path to the same level
is important [4]. The paper [5] indicated that the power
conservation is a critical issue for MANET. Some spanning
tree protocols were compared by simulations.
We propose a new protocol, BPMR (Battery Power
Management Protocol), that achieves the viability and
maximizing the number of alive devices. BPMR has the both
of the characteristics of avoiding low battery devices by
MBCR and avoiding long participation duration devices. The
route selection by BPMR avoids including low battery
devices in the route. In addition, BPMR considers
participation duration time in MANET. Thus the devices can
prolong their battery life and the number of available nodes is
kept high in the network. BPMR selects the most suitable
path as a route from a source device to a destination node
among the paths between the source node and the destination
node. There are proactive type protocols, reactive type
protocols, and hybrid type protocols for MANET routing
protocols. Though we consider reactive type routing
protocols in this paper, the idea of BPMR can be applied to
the other types of routing protocols. We compare BPMR with
the other protocols by simulations. BPMR keeps the
availability of devices high for a longer time than the other
routing protocols.
The rest of this paper is organized as follows. Section II
shows current routing protocols for MANET. Section III
shows problems in current routing protocols. In Section IV,
we show the detail of the proposed protocol and the results of
the simulations. We conclude the paper in Section V.
II. CURRENT ROUTING PROTOCOLS FOR MANET
A. DSR (Dynamic Source Routing)
DSR [8] is a reactive type protocol that constructs
communication paths on-demand. The routes are requested at
the time when the communication is necessary. The route
selection method is as follows. Note that in the rest of the
paper, a wireless communication device is denoted by a node
xi. We show the detail of the path selection algorithm by
DSR.
1) The source node generates a RREQ packet and the
packet is broadcasted to the neighboring nodes in its
effective radio wave range.
2) A node that receives the RREQ packet checks whether
its destination node is itself. If it is not the destination, it
writes its node ID to the RREQ packet and broadcasts
the RREQ packet.
3) If a node that receives the RREQ packet is the
destination node, it sends a RREP packet to the source
node. A RREP packet is sent from the destination node
to the node from which the RREQ packet arrived. The
node that receives the RREP packet sends the RREP
packet to the node from which the corresponding RREQ
packet arrived. By repeating this procedure, a RREP
packet is sent from the destination node to the source
node through the reverse of the path that the RREQ
packets are sent.
4) The source nodes select the route from which the first
RREP packet arrived to the source node.
B. MBCR (Minimum Battery Cost Routing)
MBCR [7] adjusts the nodes’ battery levels to improve the
availability of MANET. MBCR is a routing path selection
algorithm and can be used with any current routing protocols.
At the path selection, MBCR selects the path with the
maximum total remaining battery on the path. Node xi’s
current remaining battery level (1-100) is denoted by BL[xi].
The evaluation value of a path P = (x1, x2,..., xn) from x1 to xn,
PathValue(P), is
PathValue(P)=
1
1
[ ]
n
i iBL x
(1)
where xi (2 ≤ i ≤ n−1) are the relay nodes. In the above
equation, the evaluation value of a node is low if its battery
level is high. For a set of paths obtained by RREPs, MBCR
selects the following SelectPath as the route from x1 to xn.
SelectPath = argmin{PathValue(P)|P A}, (2)
where A is the set of paths obtained by RREPs.
We show the detail of the path selection algorithm by
MBCR. This algorithm inherits DSR thus we indicate
additional and different parts from DSR algorithm.
1) The same as DSR.
2) A node that receives the RREQ packet checks whether
its destination node is itself. If it is not the destination, it
writes its node ID and its remaining battery level (BL)
to the RREQ packet and broadcasts the RREQ packet.
3) The same as DSR.
4) The source node selects the route using (1) and (2) to the
received RREP packets.
III. PROBLEMS IN CURRENT ROUTING PROTOCOLS
DSR selects the minimum delay path at the path selection.
It means that DSR tends to select the minimum hop path at
the path selection. The most suitable node for relaying is
always selected by DSR as a relay node. Thus, it is unfair that
the amount of relay packets is different among nodes. Nodes
might not participate in the MANET all the time. Some nodes
might participate in MANET when they need to
communicate with the other nodes and leave from MANET
immediately after they finish their necessary communication.
MANET can be established only when there are many
volunteer participation nodes. Thus, quickly-leaving nodes
are not beneficial for the entire MANET. On the other hand,
some nodes might keep to participate to MANET for a long
duration and forward many packets for the other nodes. Such
nodes contribute very much to the entire MANET. In the
point of view of fairness, the nodes that have contributed to
Journal of Advances in Computer Networks, Vol. 4, No. 1, March 2016
14
the MANET must not be selected as a relay node. In current
routing protocols, the possibility to be selected as a relay
node is the same between a newly participated node and a
node that has relayed many packets for the other nodes. We
introduce this type of fairness to MANET routing protocols.
In addition, the number of participating nodes is important
to obtain a dense network topology. When the battery is
exhausted in a node, the node falls out from the MANET and
the number of participating nodes decreases. It is not
desirable for MANET. Therefore it is necessary to avoid
falling out nodes by battery exhaustion. MBCR avoids falling
out nodes and achieves high availability of MANET. In order
to indicate this characteristic of MBCR, we executed a
simulation to compare DSR and MBCR by self-made
simulator using C-language. The simulator locates 50 nodes
in 500m × 500m space and generates random communication
requests. The moving speed of nodes is 4km/h in average.
The nodes’ maximum battery capacity is 5550mWh
(1500mAh, 3.7V). Each node’s initial battery level is
randomly selected between 1% and 100%.
Fig. 1. Comparison of DSR and MBCR with number of alive nodes.
Fig. 1 shows the change of the number of alive nodes of
DSR and MBCR by the simulation. In Fig. 1, MBCR
decreases the number of falling out nodes more than DSR.
Since MBCR considers the network availability, there are
cases that the power consumption increases. MBCR does not
select a path in which there is a very low battery level node
and uses another path even if its number of hops is not the
minimum. It results in increasing the total power
consumption.
Fig. 2. Comparison of DSR and MBCR with sum of remaining battery
capacity of all nodes.
Fig. 2 shows the change of the sum of remaining battery
among alive nodes by the above simulation. In Fig. 2, the
sum of every node’s battery capacity is lower by MBCR than
the one by DSR. Both of avoiding falling out nodes and
keeping large number of alive nodes are important for high
availability. However excessive protection of low battery
nodes increases the total power consumption in entire
MANET. It might result in falling out many nodes. Thus
routing protocols must take a balance between protecting low
battery nodes and lowering the total power consumption. As
shown in Fig. 2, MBCR might protect low battery nodes too
much. We show that by introducing the participation duration,
such excessive protection can be avoided.
IV. THE PROPOSED PROTOCOL
A. The Need for the Proposed Protocol
We propose BPMR (Battery Power Management Routing)
protocol that keeps large number of alive nodes by
considering the participation duration. MANET protocols
must consider the limited battery in each node. It must keep
more number of nodes alive and construct a dense network
topology. However, battery consumption is not the only
characteristic to be achieved. Some nodes might leave from
MANET immediately after it has finished its necessary
communication. These nodes are not beneficial for entire
MANET. In contrast, some nodes keep to participate to
MANET for a long duration even if it has no communication
request. Such a node has contributed to entire MANET. The
nodes in MANET voluntarily relay the other nodes’ packets.
Thus, if the protocol is not fair among nodes, the incentive to
join MANET is low and few nodes join to MANET. It results
in a sparse (or disconnected) MANET topology and poor
availability. In order to achieve fairness, this paper proposes
BPMR that considers participation duration. In addition,
considering fairness improves the availability because the
load is balanced among nodes, the power consumption is also
balanced among nodes, and thus the number of falling out
nodes decreases. BPMR modifies the path selection
algorithm in routing protocols, thus it can be used with any
routing protocols. In this paper, we show modifying DSR as
an example.
B. Fairness Problem of Relay Burden
In DSR, the probability to be selected as a relay node does
not depend on the participation duration. Before we show the
detail of BPMR, we describe SDP (Simple Duration
Protocol) that is the main idea of BPMR. SDP selects the
relay nodes by considering the participation duration. The
participation duration (PD) in the MANET is a criteria that
represents the amount of contribution done as a relay node.
Some nodes might not have forwarded so many packets
during its participation duration, but the participation
increased the potential of successful communication, thus we
use not the number of forwarded packets but the participation
duration as the criteria. The node of short participation
duration can be selected as a relay node more than a longer
participation duration node. We propose the evaluation
equation of each path in SDP as follows.
PathValue(P)=1
2
1(1 )
[ ]1
32
n
iiPD x
(3)
where a path in RREP P=(x1, x2,..., xn), x1 is the source node,
xn is the destination node, xi(2 ≤ i ≤ n−1) are the relay nodes,
Journal of Advances in Computer Networks, Vol. 4, No. 1, March 2016
15
and PD[xi] is xi’s participation duration (0.0 ≤ PD[xi]).
SelectPath = argmin{PathValue(P)|P A}, (4)
where A is the set of paths obtained by RREPs. A node starts
recording the participation duration if one of the following
events occurs.
1) The node broadcasts a RREQ packet for its first
communication request.
2) The node receives a RREQ packet for the first time.
The coefficient value 32 for PD in (3) is obtained as the
best value by a simulation. Though the most suitable
coefficient of the term of PD depends on the number of nodes,
the frequency of participating and leaving, and so on, it does
not change so much from 32 in every situation. (3) and (4)
mean that the possibility that short participation duration
nodes are selected as a relay node is high and the possibility
of long participation duration is low. SDP protocol protects
long participation duration nodes from battery exhaust. Thus
SDP improves the availability of MANET. We show the
effectiveness of the proposed protocol by simulations.
We execute the following simulation to show the
characteristic of proposed SDP protocol. 50 nodes are
randomly located in 500m 500m space and random
communication requests are generated. The moving speed of
each node is randomly selected so that the speed is 4km/h in
average. The nodes’ maximum battery capacity is 5550mWh
(1500mAh, 3.7V). Each node’s initial battery level is
randomly selected between 1% and 100%. We add the
following settings for the node participation and exit. In
every minute, one randomly selected node exits from
MANET and one new node participates to MANET. There
are some long duration nodes in MANET. In order to locate
long participation duration nodes, initially selected 25 nodes
never exit from MANET.
Fig. 3. Comparison of DSR, MBCR, and SDP with the number of alive
nodes.
Fig. 3 indicates the number of alive nodes of DSR, MBCR,
and SDP. The number of alive nodes is the highest in MBCR
before 9000 sec. This is because MBCR does not select low
battery nodes as a relay node. The number of alive nodes is
almost the same in DSR and SDP until 4000 sec, because
these protocols do not have characteristics of avoiding low
battery nodes. However the number of alive nodes is the
highest in SDP from about 9000 sec. It means that MBCR
excessively protects low battery nodes by their neighbor
nodes with high battery. Thus, MBCR decreased the sum of
remaining battery in all nodes than SDP. It results in the
situation that many nodes exhaust their battery after 9000 sec.
Fig. 4. Comparison of DSR, MBCR, and SDP with sum of the remaining
battery capacity of all nodes.
Fig. 4 shows the amount the sum of remaining battery
capacity of all nodes by the above simulation. The sum of
remaining battery is the highest by SDP among three
protocols. Since SDP considers participation duration in the
evaluation equation, short participation duration nodes are
selected as a relay node more than long duration participation
nodes. Such nodes’ battery capacities are hard to decrease.
Thus, the remaining battery of long participation duration
nodes does not decrease within a short period.
However, SDP protocol might select a low battery node as
a relay node. Thus some low battery nodes exhaust battery
and fall out from MANET and the number of alive nodes
decreases. It might result in sparse network topology, few
alternative routes for communication, and decreasing the
number of reachable nodes. Thus SDP needs to incorporate
the characteristics of MBCR that avoids the nodes of low
battery level.
C. Path Selection Algorithm of BPMR
SDP achieves high remaining battery and thus the number
of alive nodes is large when many nodes fall out, but the
number of alive nodes is not the best when few nodes fall out.
In contrast, MBCR achieves the number of alive nodes large
while few nodes fall out. The characteristics of MBCR is
very important to maintain alive node ratio. Thus, we
incorporate characteristics of MBCR and SDP to BPMR
protocol. We modify the path selection algorithm of SDP to
consider remaining battery level (BL) as in MBCR.
PathValue(P)=1
2
1 1(1 )(1 )
[ ][ ]1
32
n
ii i
v vPD xBL x
(5)
where a path in RREP P=(x1, x2,..., xn), x1 is the source node,
xn is the destination node, xi (2 ≤ i ≤ n−1) are the relay nodes,
PD[xn] is xn’s participation duration (0.0 ≤ PD[xn]), v is a
value representing the weight (0 ≤ v ≤ 1) of the first term and
the second term.
SelectPath = argmin{PathValue(P) | P A}, (6)
where A is the set of paths obtained by RREPs.
We show the detail of the path selection algorithm by
BPMR. Each node has a value, called Battery Level (BL),
which is from 0.0 to 1.0 that indicates its remaining battery
(0.0 means 0% and 1.0 means 100%). In addition, each node
has a value called Participation Duration (PD), which is more
than 0.0 that indicates its participation duration (The unit is
Journal of Advances in Computer Networks, Vol. 4, No. 1, March 2016
16
minutes, that is 2.0 means the participation duration is 2
minutes).
1) The same as DSR.
2) A node that receives the RREQ packet checks its
whether its destination node is itself. If it is not the
destination, it writes its node ID, the remaining battery
level (BL), and participation duration (PD) to the RREQ
packet and broadcasts the RREQ packet.
3) The same as DSR.
4) The source node selects the route using (5) and (6) to the
received RREP packets.
BPMR has the following two characteristics. If the first
term (characteristics of MBCR) of (5) is dominant, the power
consumption increases. If the second term (characteristics of
SDP) of (5) works too much, the number of the falling out
nodes increases. The second term effects keeping remaining
battery in the entire MANET. Therefore it is necessary to find
the most suitable ratio of these two terms. We propose the
best value of parameter v by a simulation. The simulation
settings is same as the previous one.
Fig. 5. Number of alive nodes for different values of the ratio v.
Fig. 5 shows the number of alive nodes for different values
of the ratio v. In Fig. 5, the number of alive nodes is the
largest when v=0.9. It means that the weight of the first term
(MBCR) is 90% and the weight of the second term (SDP) is
10%. BPMR selects relay nodes by considering participation
duration. Thus early exiting node executes many packet
relays. When MANET is operated for a long time, there are
many long participation duration nodes. MBCR does not
protect such nodes compared with SDP, thus the remaining
battery of these nodes decrease. Therefore, the number of
alive nodes by MBCR is smaller than the one by BPMR.
Thus avoiding these long participation duration nodes results
in increasing the number of alive nodes. By BPMR, the nodes
have more power resource than by the other protocols from
this characteristic.
Fig. 6. Sum of remaining battery in all nodes for different values of the ratio
v.
Fig. 6 shows the sum of the remaining battery in all nodes
for different values of the ratio v by the above simulation.
The sum of the remaining battery in all nodes is the largest
when v=0.3. However, v = 0.3, v = 0.7, and v = 0.9 achieve
similar values and the difference is very small. Compared
BPMRs (v=0.3, 0.7, 0.9) with MBCR (v=1.0), BPMR
achieves better values than MBCR. Thus, we propose the
most preferable value of v is 0.9 in BPMR because of the
number of alive nodes is the largest when v=0.9 in Fig. 5.
However, this value can be changed, if a user prefers the
number of alive nodes or the sum of remaining battery in all
nodes in MANET.
Fig. 7. Comparison of DSR, MBCR and BPMR with number of alive nodes.
We compare DSR, MBCR and BPMR (v=0.9) by a
simulation. Fig. 7 shows the number of nodes for these three
protocols. The simulation settings is same as the previous one.
The number of alive nodes is almost the same between
MBCR and BPMR until 8000 sec. From 8000 sec to 12000
sec, the number of alive nodes by BPMR is better than the
one by MBCR. Thus, for the number of alive nodes, BPMR is
the best algorithm for the entire period.
Fig. 8. Comparison of DSR, MBCR and BPMR with sum the remaining
battery capacity of all nodes.
Fig. 8 shows the sum of the remaining battery in all nodes
for these protocols by the above simulation. In Fig. 8,
compared with the other protocols, the remaining battery is
the largest in BPMR among the three protocols. This is
because BPMR reduces the power consumption of the long
duration node with participation duration (PD). BPMR
achieves larger remaining battery than the other protocols
from 4000 sec. In BPMR, the task to relay packets is assigned
to short participation duration nodes. The other protocols do
not have such a feature. Thus the remaining battery of long
participation duration nodes is larger by BPMR than the other
protocols and the sum of remaining battery is larger than the
other protocols.
Journal of Advances in Computer Networks, Vol. 4, No. 1, March 2016
17
V. CONCLUSION
This paper proposed BPMR, a new routing protocol for
MANET, in which the route selection is modified from DSR.
BPMR uses each node’s battery level and participation
duration. Long participation duration nodes are not selected
as a relay node, thus the fairness among nodes can be
achieved. In addition, this characteristics helps increasing the
remaining battery. Even if some nodes have little remaining
battery, it does not affect the possibility of MANET
communication if the nodes exit from MANET within a short
period. By selecting short participation duration nodes as a
relay node, fairness among nodes can be achieved and the
number of alive nodes increases. When a node participates
for a long duration, the probability to be selected as a relay
node decreases by BPMR. It means that the node is motivated
to participate MANET for a long duration. We showed the
effectiveness of BPMR by simulations. There is another
fairness problem in MANET that is communication amount
and frequency within a certain period of time. There are
frequent communication request nodes and infrequent
communication request nodes. Thus, we think these nodes
should not be equally treated in terms of fairness. Achieving
this kind of fairness in MANET is our further study.
REFERENCES
[1] E. M. Royer and C. K. Toh, “A review of current routing protocols for
ad-hoc mobile wireless networks,” IEEE Personal Communications
Magazine, vol. 6, no. 2, pp. 46-55, 1999. [2] R. Verma, A. Prakash, R. Tripathi, and N. Tyagi, “Throughput
enhancement of multi-hop static Ad-hoc networks through concurrent
transmission,” in Proc. 1st IEEE International Conference on Computational Intelligence on Computational Intelligence,
Communication Systems and Networks, 2009, pp. 482-485.
[3] S. Islam, N. Hider, T. Haque, and L. miah, “An extensive comparison
among DSDV, DSR and AODV protocols in MANET,” International
Journal of Computer Applications, vol. 15, no. 2, pp. 22-24, 2011. [4] C. Yu, B. Lee, and H. Youn, “Energy efficient routing protocols for
mobile Ad Hoc networks,” Wireless Communication and Mobile
Computing, Wireless Com. Mob. Computing, pp. 959-973, 2003. [5] Q. Dai and J. Wu, “Computation of minimal uniform transmission
power in Ad Hoc wireless networks,” in Proc. International
Conference on Distributed Computing Systems Workshops, 2003, pp. 680-684.
[6] R. Ramanathan and R. Rosales-Hain, “Topology control of multihop
wireless networks using transmit power adjustment,” in Proc. 9th Annual Joint Conference of the IEEE Computer and Communications
Societies, 2000, vol. 2, pp. 404-413.
[7] C. K. Toh, “Maximum battery life routing to support ubiquitous mobile computing in wireless ad hoc networks,” IEEE Communication
Magazine, vol. 39, no. 6, pp. 138-147, 2001.
[8] B. Johnson and D. Maltz, “The dynamic source routing protocol for mobile Ad Hoc network,” RFC 4728, 2003.
Masaru Yoshimachi was born in 1992. He received his bachelor of informatics degree from Kogakuin
University, Japan, in 2015.
Currently, he is a master course student in the Department of Computer Science, Graduate School of
Kogakuin University.
Mr. Yoshimachi's research interests include networks, MANET, and routing algorithms.
Yoshifumi Manabe was born in 1960. He received
his B.E., M.E., and Dr.E. degrees from Osaka University, Osaka, Japan, in 1983, 1985, and 1993,
respectively.
From 1985 to 2013, he worked for Nippon Telegraph and Telephone Corporation. From 2001 to
2013, he was a guest associate professor of Graduate
School of Informatics, Kyoto University. Since 2013, he has been a professor of the Faculty of Informatics,
Kogakuin University, Tokyo, Japan. His research interests include
distributed algorithms, cryptography, game theory, and graph theory. Dr. Manabe is a member of ACM, IEEE, IEICE, IPSJ, and JSIAM.
Journal of Advances in Computer Networks, Vol. 4, No. 1, March 2016
18